AD-774 288

A SIMPLE RESPIRATORY GUIDE FOR RESPIRA-

TORY THERAPY IN THE TRAUMA PATIENT

Michael A. Goldfarb, et al

Edgewood ArsenalAberdeen Proving Ground, Maryland

January 1974

DISTRIBUTED BY:

National Technical Information ServiceU. S. DEPARTMENT OF COMMERCE5285 Port Royal Road, Springfield Va. 22151

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE ("Iatn Data Entered)

READ INSTRUCTIONSREPORT DOCUMENTATION PAGE BEFORE COMPLETING FORM

". REPORT NUM'ER J3OVT ACCESSION NO. 3. RECIPIENT'S CATALOG HUM ER

EB-TR-.73061 "*U 7 U.If4. TITLE (and SubtPleE . 5. TYPE OF REPORT & PERIDS COVERED

GUIDE RESPIRATORY Technical ReportA SIPLERESIRATRY UID FORRESIRAORY April to October 1973THERAPY IN THE TRAUMA PATIENT

S PERFORMING ORG. REPORT NUMBER

7. AUTHOR(*) S CONTRACT OR GRANT NUMBER(s)

Michael A. Goldfarb, Terrence F. Ciurej, William J. Saccoand M. A. Weinstein

S. PERFOrtMING ORGANIZATION NAME AND ADDRESS I0 PROGRAM ELEMENT. PROJECT, TASKARE,. • WORK UNIT NUMBERS

Commander, Edgewood Arsenal Projects I E762708A090 (EA);Attn: SAREA-BL-B 30B73 (LWL);Aberdeen Proving Ground, Md. 21010 DAAD0573C0032 (AMSAA)

11. CONTROLLING CFFICE NAME AND ADDRESS 12. REOORT DATE

Commander, Edgewood Arsenal January 1974Attn: SAREA-TS-R 13. NUMBER OF PAGES

Aberdeen Proving Ground, MD 21010 .,-8-/"14. MONITCRING AGENCY NAME & AODRESS(It different Irm, Controlli.,ij Office) 15. SECURITY CLASS. (of this report)

UNCLASSIFIED15s. DECL ASSI FICATION/ DOWNGRADING

SCHEDULE

16. DISTRIBUTION STATEMENT (of thli Report)

Approved for public release; distribution unlimited.

17. DISTRIBUTION STATEMENT (of the abstract entered In Block 20, if different from Report)

Reproduced by

NATIONAL TECHNICAL.NFORMA.-,, TION SERVICE

1, 5 O.por*ment of Commerce

SP,-nf,eltf VA 22151

III. SUPPLEMENTARY NOTES

Automation of materiel concepts; design engineering and testing. Development of lightweight

garment. Research study of trauma patient data.

19. KEY WORDS (Continue on reverse *ide it neceseary and identify by block number)

Respiratory index Survival pobabilityPulmonary shunt Information gainAlveolar-arterial oxygen difference

20. ABSTRACT (Continue an, reversde sid f I noc eeisty end identif%, by block numoer)

The respiratory index (R.I.), PAaDO-PaD 2 was investigated in 177 intubated patients

treated at the Maryland Institute of Emergency Medicirie. An R.I. of 0. 1-0.37 is normal. An R.I. of

2 or greater was an indication for intubation, and over 6 was associated with a 12% probability of

survival. The R.I. reflects the presence of pulmonary shunting in a variety of circumstances

including atelectasis, pulmonary contusion or pulmonary emboli. A nomogram which allows one to

follow the course of the patient with respiratory problems is described. Movement along the same

DOD F¶OA. . 1473 EDITION OF I NOV 6S IS OBSOLETEI UNCLASSIFIED

SECURITY CLASSIFICATION OF THIS PAGE (lioen Date Nntered)

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAE(W••n Daf, Enre,")

isobars or between isobars can be followed by plotting the Pao- against the F 10 2 Thus, the

rationale and effect of respiratory therapeutic manipulations may be graphically recorded.

UNCLASSIFIED

2 SECURITY C.ASSIFICATION OF THIS PAGE(Whan Da.la Ent$rad)

SUMMARY

The respiratory index (R.I.), PAaDO 21/PaO 2 was investigated in 177 intubated patients

treated at the Maryland Institute for Emergency Medicine. An R.I. ol 0. 1-0.37 is normal. An R.l. of2 or greater was an indication for intubation, and over 6 was assotiated with ,1 12% probability ofsurvival. The R.I. reflects the presence of pulmonary shunting in a variety of circumstancesincluding atelectasis, pulmonary contusion or pulmonary emboli.

A nomogram which allows one tc follow the course of the patient with respiratoryproblems is described M11ovement along the same isobars or between isobars can be followed byplotting the PaO) against the Flo,. Thus, the rationale and effect of respiratory therapeuticmanipulations may be graphically recrded.

PREFACE

The work described in this report was authorized under Projects I E762708AO90(EA)Automation of Materiel Concepts-Design, Engineering and Testing: 30B73(LWL) Development ofLightweight Garment: and DAAD0573CO032(AMSAA) Research Study of Trauma Patient Data.This work was started in April 1973 and completed in October 1973.

Reproduction of this document in whole or in part is prohibited except withpermission of the Commander, Edgewood Arsenal, Attn: SAREA-TS-R, Aberdeen Proving Ground,Maryland 21010; however, DDC and the National Technical Information Service are authorized toreproduce the document for US Government purposes.

Acknowledgments

We wish to thank Dr. R. A. Cowley of the Maryland Institute for Emergency Medicine,Baltimore, Maryland, for allowing us to utilize their computerized data bank and patients' records.We also wish to thank Mr. Paul H. Broome for the computerization of the data herein and Ms.Marion Royston for technical assistance in the preparation of this report: both of whom are fromBiomedical Laboratory, Edgewood Arsenal, Maryland.

3

CONTENTS

Page

1. INTRODUCTION ............. ........................ 5

11. METHOD ................ ........................... 6

111. RESULTS AND DISCUSSION .......... ................... 6

IV. RESPIRATORY INDEX NOMOGRAM ....... ................ 9

V. FUTURF EFFORTS ............. ....................... I 1

LITERATURE CITED ........ ..................... .... 13

GLOSSARY ................ .......................... 15

APPENDIX ........ ............................ ... 17

DISTRIBUTION LIST ................................ ... 19

4

A SIMPLE RESPIRATORY GUIDE FOR RESPIRATORY THERAPYIN THE TRAUMA PATIENT

I. INTRODUCTION.

The purpose of this research is to test a respiratory index (R.I.) as an indicator of atrauma patient's respiratory state. If a patient's respiratory state could be simply characterized and

followed, this would allow one to: (I) compare therapy in patients with respiratory complicationsin various institutions, (2) compare variations in treatment, and (3) graphically represent a patient'sprogress or deterioration as an adjunct to patient care by means of a nomogram described later.

An increase of the alveolar-arterial P 2 difference is a common indicator of hypoxiaP2and is an important consideration in controlling arterial oxygenation in the clinical environment.The alveolar-arterial PO,, difference results from venous admixture (or physiologic shunt) which is

caused by (a) shunted venous blood which mixes with oxygenated blood leaving the pulmonarycapillaries, and (b) uneven ventilation-perfusion ratios in different parts of the lung., 2

When the patient breathes 100% oxygen the PAO 2 may be calculated from:

AO2 = PB + PH20,T) = (PB-PHT) X 1.0- PaCO2

where

PB= barometric pr,,ssure

PH 20,T = alveolar water vapor pressure at the patient's temperature (T),

(approximately 47 mm Hg)F10 2 = fractional concentration of 02 in inspired gas

PaCO 2 = arterial partial pressure of carbon dioxide assumed to be equal tothe alveolar partial pressure of the carbon dioxide (PACo2). 3

The above equation permits one to calculate the alveolar-arterial oxygen difference at differentinspired oxygen concentrations, between 20% and 100%. The alveolar-arterial difference was thendivided by the PaO 2 so that the respiratory index (R.I.) equals:

[(PB- PH20,T)F10 2 -PaC02] -PaO 2 PAaD0 2

= - R.I.

PaO 2 PaO 2

This calculation was suggested to the investigators by J. H. Siegel.* 4 Since the PAaD0 2 reflects

shunting, the R.I. would also reflect pulmonary shunting. And this paper will indicate the usefulness

* Siegel, J. H. Personal Communication.

5

Iof the R.I. when a patient's respiratory pathophysiology involves shunting. These conditions includepulmonary contusion, pulmonary embolus, and atelectasis caused by bronchial obstruction,mechanical compression of the lung or hypoventilation.3

II. METHOD.

The patients evaluated here were all treated at the Maryland Institute for EmergencyMedicine. Most patients were trauma victims and were delivered by helicopter from the scene of theaccident between March 1971 and August 1972. Of the primary types of insults, 71.8% incurredblunt trauma, 7.3% gunshot wounds, 2.8% burns, 5.1% elective surgical problems, and 13.0%medical problems. Of the trauma group, with regard to major areas affected, 27.4% had centralnervous system injury, 25.4% thoracic injury, 17.9% abdominal irjury, 17.5% musculoskeletaltrauma, and 11.9% head injury. Patients listed usually had one or more of the above injuries.

The patients included here were in the unit at least two days. They represent patientswho had the Flo 2 recorded throughout their therapy. This means that the patient was intubated at

least once during his hospital course and placed on the Engstrom Respirator. The EngstromRespirator was used in every case when the patient required assisted ventilation. Since the patientswere on Engstrom Respirators, one could calculate the percent 02 administered by plotting theliters per minute of 02 against the total minute volume in liters per minute. This graph accompaniesall Engstrom Respirators. Arterial blood gases were measured every time the therapy of FIO2 was

altered. Arterial blood gases were also measured at least every 6 hours in every intubated patient.

III. RESULTS AND DISCUSSION.

A. Respiratory Index Data and Associated Probabilities of Survival.

The accompanying retrospective statistics in table I represent a total of 177 patients,where 116 lived and 61 died. The upper limits of normal alveolar-arterial differences (on room air)for ages 20,40, and 60 are 19 mm Hg, 24 mm Hg, and 28 mm Hg. The lower limits for PaO 2 (onroom air) for those age groups are 85 mm Hg, 80 mm Hg, and 75 mm Hg. Dividing thealveolar-arterial difference by the PaO2 the respiratory indices are 0.22, 0.3 and 0.37, respectively.5

In every case when the R.I. was as high as 2, the patient was intubated. This is understandable sincean R.I. of 2, with the patient on room air (20% oxygen), and a PaCO of 35 mm Hg would meanthat the patient would have a PaO2 of 37 mm Hg. The R.I.'s listed represent the maximum R.I.'s ofthe patients' entire hospital stay (table 1). The R.I. probauilities are generated from the group ofpatienis treated at the Maryland Institute for Emergency Medicine. The probabilities should beexpected to change somewhat depending upon the therapy at a particular institution.

When the pathology was reviewed, a major disorder in 51 out of 116 survivors was

central nervous system damage usually due to head trauma. Twenty nine out of 61 patients whodied also had major CNS trauma. Brain death secondary to a CNS injury was most important in the

group that died with and R.I. under 4. There were 16 patients with CNS injury among the 22 patientsthat died with an R.I. under 4. The other six patients that died with an R.I. under 4 died of

6

Table 1. Frequency Charts for Maximum Respiratory Index

Maximum Total Number Patients that Total Number Patients thatRespiratory of Patients Lived without of Patients Died withoutIndex (R. 1.) (Lived) CNS Injury (Died) CNS Injury

(0-1) 33 19 2 0

(1-2) 23 15 6 2

(2-3) 21 12 8 1

(3-4) 16 8 6 3

(4-5) 12 6 7 6

(5-6) 8 2 7 5

(6-7) 2 2 5 3

(7-8) 6 5

(8-9) 5 1

(9-10) 3 3

(10-11) 3 2

(11-17) 3 1

TOTAL 116 65 61 32

7

I1

sepsis and/or cardiac arrhythmias. Respiratory pathology did not appear to be a major factor intheir demise. The patients with CNS injury that died with a low R.I. (under 2) had a median rime todeath of 5 days. These patients did not enter with, nor develop, clinically significant respiratorycomplications. The patients with CNS injury and an R.I. from 2 to 4 had a median time to death of10 days. They also died primarily a orain death, but frequently had intercurrent pneumonia orassociated thoracic or abdominal injury, leading to pulmonary shunting and a higher R.I.

Of the entire group of patients that died with an R.I. over 5, pulmonary pathology wasalways an important element in their demise. There was only oiie patient that lived with an R.I.over 7. This patient had an R.I. of 10.5 for a bnef time subsequent to a bilateral pneumothoraxsecondary to positive and expiratory pressure (PEEP). Once the pneumothoraces were treated, thepatient's R.I. sr ,wly returned to normal. Eight patients survived with an R.I. between 5 and 6, andonly twi survived with an R.I. between 6 and 7. Thus, II patients survived with an R.I. over 5. Onthe other hand, 32 of the 61 patients that died had an R.I. of 5 or greater. The probability ofsurvival with an R.I. over 5 was 26% and an R.I. over 6 was associated with a 12% probability ofsurvival. Other survival probabilities are listed in table II. It should be stressed that the reliability ofthe R.I. is enhanced by its irreversibility. For example, even if the R.I. is over 6 and then returns tonormal with treatment over the next several days, the patient appears to have the same poorprobability of survival. In addition, in one-half of the patients with an R.I. greater than 5 theirhighest measurements were recorded at least 1 day prior to the day of death.

Table II. Histograms for Maximum Respiratory Index

Maximum Probability of Survival Probability of SurvivalRespiratory for Total Number For Patients withoutIndex (R.I.) Of Patients CNS Injury

(0-1) 0.95 0

(1-2) 0.8 0.89(2-3) 0.73 0.93

(3-4) 0.73 0.73

(4-5) 0.64 0.49(5-6) 0.55 0.28>5 0.26 0.2

>6 0.12 0.18

InformationGain 0.20* 0.29*

* See Appendix

8

IV. RESPIRATORY INDEX NOMOGRAM.

In order to track respiratory therapy graphically for this group of 177 patients, anomogram was designed(figure). Respiratory index isobars were calculated with varying Pao andpercent 02 administered, but with a constant PaCO2 of 35 mm Hg. The variations in R.I. ;ith aPaCO2 of 20 mm Hg or 60 mm Hg instead of 35 mm Hg have been calculated. For example, if the

patient is on 40% 02, and the PaO- is 100 mm Hg, and the PaCO2 ranges from 20 mm Hg to235 mm Hg to 60 mmHg the corresponding R.I.'s are 0.3, 0.15, and 0.10. This variation decreases asthe R.I. increases. For a closer approximation, one can add 0.2 R.I. units for every 15 mm Hgdecrease in PaCO,7 from a PaCO2 of 35 mm Hg and subtract 0.2 R.I. units for every 15 mm Hgincrease in PaCO 2 "ibove 35 mm Hg.

In order to explain the use of this graph consider a typical patient with thoracictrauma. The patient is a young adult who incurred fractures of ribs, 3, 4, 5, 6, 7, and 8 on the leftside. This represents an example of blunt chest trauma with a lung contusion. Twenty minutes afterthe accident the PaO is 50 mm Hg, 35 mm Hg, and arterial blood pH 7.3 while on room air(point A). The graph indicates that the patient is on an R.I. isobar of about 1. The patient thenbegins to have respiratory muscular splinting and is breathing at 20 times per minute. He is given30% oxygen by mask. Repeat measurements of blood gases show a PaO2 of 55 mm Hg, P

35 mm Hg, and arterial blood pH 7.4; the patient moves to the 2.0 R.I. Isobar (point B). At thispoint the patient should 1e intubated. In this example he is placed on 40% oxygen on a respirator.The PaO is now 120 ram Hg and the R.I. is again 1 (point C). He has, therefore, moved from anR.I. of 2 to 1, indicating an improved respiratory status secondary to therapy. If a repeat PaO2 is205 mm Hg, the patient would be on 0.2 isobar (point D), indicating further improvement. If repeatPaO , however, is 58 mm Hg, the patient's condition would have deteriorated and he would now beon t"ie 3.0 R.I. isobar (point E). In the instance when the Pao 2 is 205 mm Hg, one might place thet a 0 2

patient on 20% 02 on a respirator which should give a PaO2 of 90 mm Hg, assuming the patient

remains on the 0.2 R.I. isobar (point F). If the PaO2 drops to 58 mm Hg, a change of therapy isobviously indicated. One course might be increasing the administered 02 to 60% which should

give a PaO2 of 90 mm Hg (point G). If the PaO2 has not increased to at least 90 mm Hg on 60% 02

(indicating an R.I. of 3), end expiratory pressure should be added to the regimen. If the PaO2

drops from 90 mm Hg to70 mm Hg on a 60% 02, PEEP would also be indicated. It should be notedthat there is miminal increase in PaO once a patient is on the R.I. 3 isobar. Here, if the patient ison 60% 02 and the Pao is 90 mm I-g (point G), placin' the patient on 100% 09 would give a PaO 2of 160 mm Hg (point H?. The amount of increase in the PaO2 decreases as the lopes of the isobarsdecrease. This reflects the generally accepted notion that if a patient is sufficiently ill to require60% FIO0 , increasing the FO0 beyond 60% does not effect much of an increase in Pao 2 . In

2 2..addition, the risk of oxygen toxicity is a hazard when the patient is placed on high concentrationsfor long periods.2

The R.I. in certain cases could also inform the physician that the respirator is notworking properly. If any patient is supposedly breathing 30% 02, and his PaO2 were 220 mm Hg,

he would have an R.I. less than 0, which is impossible. This would indicate that the FIo 2 was

9I

P9O2- ARTERIAL PARTIAL PRESSURE OF OXYGEN

00 ! I-

0

I,:IC,!

in

10

m0 O0

oz

z

o 00 o o 0 "

X2aNI ASOIVHldS38

10

actually greater than 30% and probably at least 40%. In several intubated patients, therapy includedlimited periods of 100% 02administration, particularly when calculating the pulmonary shunt.Here, one noted that the typical patient remained near the same isobar as the amount of 02administered changed from 40%/o to 100% and back to 40%.

V. FUTURE EFFORTS.

There are several areas of research that require further investigation with regard to therespiratory index. They include:

1. Testing the R.I. with data from trauma patients in a prospective fashion.Subgroups of trauma patients may clarify the prognostic significance of respiratory pathology incases of isolated thoracic trauma, multiple trauma including the thorax, the multiple traumaexcluding the thorax. The relevance of the R.I. in pulmonary units (nonsurgical) should provehelpful with regard to those cases with significant shunting, and deserves implementation.

2. The applicability of the nomogram will be tested further by the physicianshandling respiratory problems as a graphic representation of the trend of various forms ofrespiratory pathology with alterations in therapy.

3. The prognostic and therapeutic value of the R.I. must be compared with morecomplex pulmonary indices including percent pulmonary shunting, pulmonary compliance, andexpiratory reserve volume, etc. Such indices may perhaps be added to the R.I. to characterize evenmore precisely and collate respiratory pathophysiology.

VI. LITERATURE CITED

1. Nunn, J. F. Applied Respiratory Physiology. pp 337-338. Butterworth, London,England. 1969.

2. Pontoppidan, H., Geffin, B., and Lowerstein, E. Acute Respiratory Failure in theAdult. New England Journal of Medicine. 287 (15), 743-752 (1972).

3. Laver, M. B. and Austen, W. G. Cardiorespiratory Dynamics. Surgery, 2d Edition,Chapter 1, pp 2-7. W. B. Saunders Company, Philadelphia, Pennsylvania. (1969).

4. Siegel, J. H. and Farrell, E. J. A Computer Simulation Model t6 Itudy the Clini-cal Observability of Ventilation and Perfusion Abnormalities in Human Shock StuCes. Surgery 73(6), 898-912. (1973).

5. Mellemgaard, K. The Alveolar-Arterial Oxygen Difference: Its Size andComponents in Normal Man. Ata Physiol. Scand., 67, 10-20 (1966).

11

GLOSSARY

PAO 2 Partial pressure, alveolar, of oxygen

PB Barometric pressure

Arterial partial pressure of carbon dioxide assumed toPaCO2 be equal to the alveolar partial pressure of the

carbon dioxide

PH2 0,T Alveolar water vapor pressure at the patient's temperature (T)approximately 47 mm Hg

F10 2 Fractional concentration of 02 in inspired gas

Pa0 2 Partial pressure, arterial, of oxygen

PAaDO 2 Alveolar-arterial partial pressure of oxygen difference

PEEP Positive and expiratory pressure

12

-- -- i

[ APPENDIX

INFORMATION GAIN

A measure of the information inherent in an index is called the "information gain". Ithas the following interpretation: Suppose an Intensive Care unit manages to save 80% of theirpatients. The probability of survival, PL, of an arrivinig patient is, therefore, 0.8. If the physician isgiven aaditional information about a particular patient, he might be able to alter the prognosis from0.8 to 0.2. The gain in information between probabilities would be the difference before and afterthe additional input. More specifically, with the total group of patients in this study, the averageinformation gain is 0.2 with respect to the lived-died statistics of intubated patients. For the group

of patients without CNS injury the average information gain is 0.29. Let x designate the index to beevaluated. Then the average information gain, I, is

nn2 P I PL- P(LIx is in bin i) j[Probability that x is in bin i]

where

PL is the prior probab;ity that a patient will live, and

P(LIx is in bin i) is the conditional probability

that a patient will live given that index x belongs to bin i.*

The information gain may also be used to test any other parameter which one may want to includeto enhance the validity of a prognosis. The information gain analysis would indicate the value of thenew parameter.

Sacco, W. J. and Copes, W. S. Reduction of the Class of Feature Evaluation Techniquesof Pattern Analysis. Pattern Recognition 4, 331-332 (1972).

13